89     float16x8_t vacc4x01234567 = vacc0x01234567;  in xnn_f16_gemm_ukernel_8x8__neonfp16arith_ld64()  local
112           vacc4x01234567 = vfmaq_lane_f16(vacc4x01234567, vb01234567c0, va4, 0);  in xnn_f16_gemm_ukernel_8x8__neonfp16arith_ld64()
130         vacc4x01234567 = vfmaq_f16(vacc4x01234567, va4c0, vb01234567c0);  in xnn_f16_gemm_ukernel_8x8__neonfp16arith_ld64()
142           vacc4x01234567 = vfmaq_lane_f16(vacc4x01234567, vb01234567c1, va4, 1);  in xnn_f16_gemm_ukernel_8x8__neonfp16arith_ld64()
160         vacc4x01234567 = vfmaq_f16(vacc4x01234567, va4c1, vb01234567c1);  in xnn_f16_gemm_ukernel_8x8__neonfp16arith_ld64()
172           vacc4x01234567 = vfmaq_lane_f16(vacc4x01234567, vb01234567c2, va4, 2);  in xnn_f16_gemm_ukernel_8x8__neonfp16arith_ld64()
190         vacc4x01234567 = vfmaq_f16(vacc4x01234567, va4c2, vb01234567c2);  in xnn_f16_gemm_ukernel_8x8__neonfp16arith_ld64()
202           vacc4x01234567 = vfmaq_lane_f16(vacc4x01234567, vb01234567c3, va4, 3);  in xnn_f16_gemm_ukernel_8x8__neonfp16arith_ld64()
220         vacc4x01234567 = vfmaq_f16(vacc4x01234567, va4c3, vb01234567c3);  in xnn_f16_gemm_ukernel_8x8__neonfp16arith_ld64()
245         vacc4x01234567 = vfmaq_f16(vacc4x01234567, va4, vb01234567);  in xnn_f16_gemm_ukernel_8x8__neonfp16arith_ld64()
259     vacc4x01234567 = vmulq_f16(vacc4x01234567, vscale);  in xnn_f16_gemm_ukernel_8x8__neonfp16arith_ld64()
269     vacc4x01234567 = vminq_f16(vacc4x01234567, vmax);  in xnn_f16_gemm_ukernel_8x8__neonfp16arith_ld64()
279     vacc4x01234567 = vmaxq_f16(vacc4x01234567, vmin);  in xnn_f16_gemm_ukernel_8x8__neonfp16arith_ld64()
293       vst1q_f16(c4, vacc4x01234567);  in xnn_f16_gemm_ukernel_8x8__neonfp16arith_ld64()
317       float16x4_t vacc4x0123 = vget_low_f16(vacc4x01234567);  in xnn_f16_gemm_ukernel_8x8__neonfp16arith_ld64()
335         vacc4x0123 = vget_high_f16(vacc4x01234567);  in xnn_f16_gemm_ukernel_8x8__neonfp16arith_ld64()